Install & First Run¶

This page gets you from nothing to a working AI response in a few minutes.

1. Requirements¶

Python 3.10 or newer
A connection to an LLM provider. The simplest is local Ollama (free, no API key), but any OpenAI-compatible provider works.

pip install hands-on-ai

Or install the latest from GitHub:

pip install git+https://github.com/michael-borck/hands-on-ai.git

That single command installs everything: all four modules, plus PDF and Word support for RAG. There are no extra installs to remember.

The quickest path is local Ollama (no API key):

# Install Ollama from https://ollama.com, then pull a model:
ollama pull llama3

Ollama runs at http://localhost:11434 and Hands-On AI connects to it automatically, with no configuration needed.

Using OpenAI, OpenRouter, Groq, an authenticated Ollama server, or anything else? See Choose a Provider.

from hands_on_ai.chat import get_response

print(get_response("Explain photosynthesis like I'm 10."))

You should see a real response from the model. Try a personality bot:

from hands_on_ai.chat import pirate_bot

print(pirate_bot("Tell me about sailing ships."))

Hands-On AI ships a diagnostic command that checks your connection, shows the resolved configuration, and lists available models:

handsonai doctor

If something isn't connecting, this is the fastest way to see why (wrong server URL, missing API key, model not pulled, etc.).